Extracting qualitative knowledge from large quantities of data

Abstract:
Extracting qualitative knowledge from large quantities of data is just the beginning of our search for meaning and a plausible explanation for complex phenomenons (from urban dynamics to analysis and prediction of human behavior on social networks). The focus of our research has been identifying the sound interplay between different phenomenological and contextual features used for predictive modelling of the studied phenomenon. The focus of my presentation will be on our approach to urban modelling, which starts with semantically-enriched representations of urban spaces as a basis for agent-based modelling of urban dynamics. I will also present our research work in predictive social computing, which spans from personality prediction on Facebook and YouTube video monologues to real-time multimodal emotion recognition in a well-being mobile application.

BIO:
Sonja Gievska is an Associate Professor of Computer Science and Engineering at the Univ. of Ss. Cyril & Methodius, Macedonia. Her research interests lie in the area of intelligent interactive systems, with a special focus on agent-based modeling and social computing. She earned her PhD in agent-based interfaces from the Computer Science Department at The George Washington University and has continued to collaborate with this department for the past 15 years (in the roles of Research Associate, Visiting Scholar, and Adjunct Faculty). Gievska is an Associate Editor of Elsevier Computers & Education Journal.